Although some theorists (e.g., Bem, 1967, Cooley, 1902; Mead, 1934) have speculated about how people draw inferences from their own behavior and the feedback of others, the question of how self-relevant knowledge is combined to form one's self-concept has received little attention. A great deal of work on self-concept has mainly focused on the implications of already-developed self-information processing mechanisms involved in forming self-concepts. The proposed research attempts to fill this gap in our knowledge, while exploring factors that affect how self-concepts are formed. The current work draws on the on-line vs. memory-based processing distinction (Hastie & Park, 1986), which has been used to understand how people form impression about individual and group social targets (McConnell, 1999; McConnell, Sherman, & Hamilton, 1994b, 1997). It is proposed that people's expectations about the self (e.g., implicit person theories, cultural beliefs about self-construal), self-focus (e.g., self- awareness, self-monitoring), and social interaction contexts (e.g., accountability, social comparison) influence the extent to which they process self-relevant information in an integrative fashion. As expectations about the self grow more consistent and diagnostic of one's "true essence", focus on the self increases, and social contexts encourage greater self-understanding , on-line self-concepts should result. These experiments (four preliminary, ten proposed) test conditions that determine whether memory-based or on-line self concepts are produced. In addition to suggesting a process account for self-concept formation, it is predicted that the way in which one forms a self-concept (memory- based versus on-line) will have meaningful implications. Those who form on-line, instead of memory-based, self-concepts will be especially influenced by early feedback, and they may hold more stable and resilient self-concepts, and store self-relevant information in a more abstracted form. Thus, the current work provides a process account for how self-concepts are formed. Moreover, it predicts that self-concepts will formed differently based on particular situational and individual difference factors. Although this work posits that self concept formation shares much in common with social perception for other types of targets (e.g., individuals, groups), it also examines issues unique to the self (e.g., cultural theories about independent versus interdependent self-construal, social comparison as a means for self-assessment). Because self-concepts have important implications (e.g., behavior, personal outlook, social comparison, and experience of emotion) for social functioning and mental health, an understanding of how self-concepts are formed, and the consequences of forming them, is important.